Recently, parallelization/sharding has been getting quite a lot of attention, with various narratives being spun, but what I care more about now is: can the money safely come out once it's in? No matter how fast the chain is, how much the routing costs, in the end, getting stuck on cross-chain bridges, L2 exit windows, or even a contract whose permissions haven't been revoked feels like all the effort is wasted. Yesterday, I saw a bunch of people complaining that validators/miners rely more on MEV after their income drops, and the fairness of transaction ordering has become even more mysterious... To put it simply, whether retail traders get a bad deal or not is just as important as the "performance story." My approach is still old-fashioned: first run a small amount to see the slippage, depth, and rollback failures, and also check the exit paths (bridge/withdraw/unlock). Only after confirming it works do I add more. What about you?